From b7981569bf29aae0fcffb1e23a4cd2200eacdb4e Mon Sep 17 00:00:00 2001
From: cvanelteren <caspervanelteren@gmail.com>
Date: Sat, 7 Feb 2026 07:21:40 +1000
Subject: [PATCH] CI: preserve selected nodeids as JSON arrays

---
 .github/workflows/build-ultraplot.yml | 69 ++++++++++++++++-----------
 .github/workflows/main.yml            |  4 +-
 2 files changed, 42 insertions(+), 31 deletions(-)

diff --git a/.github/workflows/build-ultraplot.yml b/.github/workflows/build-ultraplot.yml
index 7c8bc25e8..cec61da39 100644
--- a/.github/workflows/build-ultraplot.yml
+++ b/.github/workflows/build-ultraplot.yml
@@ -135,7 +135,26 @@ jobs:
           echo "TEST_NODEIDS=${TEST_NODEIDS}"
           # Save PR-selected nodeids for reuse after checkout (if provided)
           if [ "${TEST_MODE}" = "selected" ] && [ -n "${TEST_NODEIDS}" ]; then
-            printf "%s\n" ${TEST_NODEIDS} > /tmp/pr_selected_nodeids.txt
+            python - <<'PY'
+            import json
+            import os
+
+            raw = os.environ.get("TEST_NODEIDS", "").strip()
+            nodeids = []
+            if raw and raw != "[]":
+                try:
+                    parsed = json.loads(raw)
+                except json.JSONDecodeError:
+                    parsed = raw.split()
+                if isinstance(parsed, str):
+                    parsed = [parsed]
+                if isinstance(parsed, list):
+                    nodeids = [item for item in parsed if isinstance(item, str) and item]
+            with open("/tmp/pr_selected_nodeids.txt", "w", encoding="utf-8") as fh:
+                for nodeid in nodeids:
+                    fh.write(f"{nodeid}\n")
+            print(f"Selected nodeids parsed: {len(nodeids)}")
+            PY
           else
             : > /tmp/pr_selected_nodeids.txt
           fi
@@ -152,26 +171,22 @@ jobs:
           python -c "import ultraplot as plt; plt.config.Configurator()._save_yaml('ultraplot.yml')"
           if [ "${TEST_MODE}" = "selected" ] && [ -s /tmp/pr_selected_nodeids.txt ]; then
             status=0
-            filter_nodeids() {
-              local filtered=""
-              for nodeid in $(cat /tmp/pr_selected_nodeids.txt); do
-                local path="${nodeid%%::*}"
-                if [ -f "$path" ]; then
-                  filtered="${filtered} ${nodeid}"
-                fi
-              done
-              echo "${filtered}"
-            }
-            FILTERED_NODEIDS="$(filter_nodeids)"
-            echo "FILTERED_NODEIDS_BASE=${FILTERED_NODEIDS}"
-            if [ -z "${FILTERED_NODEIDS}" ]; then
+            mapfile -t FILTERED_NODEIDS < <(
+              while IFS= read -r nodeid; do
+                [ -z "$nodeid" ] && continue
+                path="${nodeid%%::*}"
+                [ -f "$path" ] && printf '%s\n' "$nodeid"
+              done < /tmp/pr_selected_nodeids.txt
+            )
+            echo "FILTERED_NODEIDS_BASE_COUNT=${#FILTERED_NODEIDS[@]}"
+            if [ "${#FILTERED_NODEIDS[@]}" -eq 0 ]; then
               echo "No valid nodeids found on base; skipping baseline generation."
             else
               echo "=== Memory before baseline generation ===" && free -h
               pytest -n ${PYTEST_WORKERS} --dist loadfile --tb=short --disable-warnings -W ignore \
                 --mpl-generate-path=./ultraplot/tests/baseline/ \
                 --mpl-default-style="./ultraplot.yml" \
-                ${FILTERED_NODEIDS} || status=$?
+                "${FILTERED_NODEIDS[@]}" || status=$?
               echo "=== Memory after baseline generation ===" && free -h
               if [ "$status" -eq 4 ] || [ "$status" -eq 5 ]; then
                 echo "No tests collected from selected nodeids on base; skipping baseline generation."
@@ -213,19 +228,15 @@ jobs:
           echo "TEST_NODEIDS=${TEST_NODEIDS}"
           if [ "${TEST_MODE}" = "selected" ] && [ -s /tmp/pr_selected_nodeids.txt ]; then
             status=0
-            filter_nodeids() {
-              local filtered=""
-              for nodeid in $(cat /tmp/pr_selected_nodeids.txt); do
-                local path="${nodeid%%::*}"
-                if [ -f "$path" ]; then
-                  filtered="${filtered} ${nodeid}"
-                fi
-              done
-              echo "${filtered}"
-            }
-            FILTERED_NODEIDS="$(filter_nodeids)"
-            echo "FILTERED_NODEIDS_PR=${FILTERED_NODEIDS}"
-            if [ -z "${FILTERED_NODEIDS}" ]; then
+            mapfile -t FILTERED_NODEIDS < <(
+              while IFS= read -r nodeid; do
+                [ -z "$nodeid" ] && continue
+                path="${nodeid%%::*}"
+                [ -f "$path" ] && printf '%s\n' "$nodeid"
+              done < /tmp/pr_selected_nodeids.txt
+            )
+            echo "FILTERED_NODEIDS_PR_COUNT=${#FILTERED_NODEIDS[@]}"
+            if [ "${#FILTERED_NODEIDS[@]}" -eq 0 ]; then
               echo "No valid nodeids found on PR branch; skipping image comparison."
               exit 0
             else
@@ -236,7 +247,7 @@ jobs:
               --mpl-results-path=./results/ \
               --mpl-generate-summary=html \
               --mpl-default-style="./ultraplot.yml" \
-              ${FILTERED_NODEIDS} || status=$?
+              "${FILTERED_NODEIDS[@]}" || status=$?
               echo "=== Memory after image comparison ===" && free -h
               if [ "$status" -eq 4 ] || [ "$status" -eq 5 ]; then
                 echo "No tests collected from selected nodeids; skipping image comparison."
diff --git a/.github/workflows/main.yml b/.github/workflows/main.yml
index 176ca1a51..18a5bc6ba 100644
--- a/.github/workflows/main.yml
+++ b/.github/workflows/main.yml
@@ -74,7 +74,7 @@ jobs:
         run: |
           if [ "${{ github.event_name }}" != "pull_request" ]; then
             echo "mode=full" >> $GITHUB_OUTPUT
-            echo "tests=" >> $GITHUB_OUTPUT
+            echo "tests=[]" >> $GITHUB_OUTPUT
             exit 0
           fi
 
@@ -104,7 +104,7 @@ jobs:
           import json
           data = json.load(open(".ci/selection.json", "r", encoding="utf-8"))
           print(f"mode={data['mode']}")
-          print("tests=" + " ".join(data.get("tests", [])))
+          print("tests=" + json.dumps(data.get("tests", []), separators=(",", ":")))
           PY
           cat .ci/selection.out >> $GITHUB_OUTPUT